control-panel

plan 0082 — geographic colocation latency research

Boxes

box_id	region	connected	peers	ntp_offset_us	funds	last_seq	last_seen

Accounts

address	kind	box_id	label	balance_wei	low funds

Address

Kind

box_id (per_box only)

Label

Run Test 5 (competing-nonce race)

Race account

box_ids (comma-separated)

Lead time (ms)

Run Test 3/4 (periodic campaign)

test name

box_ids (comma-separated)

cadence_secs

duration_secs (optional)

Test 3 latency (median / p90) CSV

box_id	samples	median_ms	p90_ms

Test 5 win table CSV

box_id	wins	races

Arm Test 6 (reactive real-trigger pool)

Race-pool account

box_ids (comma-separated)

Test 6 live activity (real triggers)

One row per real trigger_tx_hash the production discovery pipeline on ≥1 box confirmed profitable. detected_by lead is NTP-corrected; reaction latency is box-local (no NTP correction needed, plan 0083's clock model).

trigger_tx_hash	detected_by (region, NTP-corrected lead)	reacted_by (box-local latency)	gross_profit_pls	status

Test 6 win table CSV

box_id	wins	races

Test 6 reaction latency (detection → fire, median / p90)

box_id	samples	median_ms	p90_ms

Test 6 reaction-latency distribution

Test 6 cross-box detection lead

trigger_tx_hash	first detected by	all sightings (box_id @ corrected_ts_ms)

control-panel

Boxes

Accounts

Run Test 5 (competing-nonce race)

Run Test 3/4 (periodic campaign)

Test 3 latency (median / p90) CSV

Test 5 win table CSV

Arm Test 6 (reactive real-trigger pool)

Test 6 live activity (real triggers)

Test 6 win table CSV

Test 6 reaction latency (detection → fire, median / p90)

Test 6 reaction-latency distribution

Test 6 cross-box detection lead

Live event feed